Molecular Oxygen Induced in-Gap States in PbS ... - ACS Publications

Sep 24, 2015 - Applied Science and Technology Graduate Program, §Department of Chemistry, and. ) Department of Materials Science and Engineering, Uni...
1 downloads 22 Views 1MB Size
Subscriber access provided by The Chinese University of Hong Kong

Article

Molecular Oxygen Induced In-Gap States in PbS Quantum Dots Yingjie Zhang, Danylo Zherebetskyy, Noah D Bronstein, Sara Barja, Leonid Lichtenstein, A. Paul Alivisatos, Lin-Wang Wang, and Miquel Salmeron ACS Nano, Just Accepted Manuscript • DOI: 10.1021/acsnano.5b04677 • Publication Date (Web): 24 Sep 2015 Downloaded from http://pubs.acs.org on September 24, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Nano is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

Molecular Oxygen Induced In-Gap States in PbS Quantum Dots Yingjie Zhang1,2†, Danylo Zherebetskyy2†, Noah D. Bronstein3, Sara Barja2, Leonid Lichtenstein2, A. Paul Alivisatos2,3,4,5, Lin-Wang Wang2, Miquel Salmeron2,4* 1

Applied Science and Technology Graduate Program, University of California, Berkeley, CA

94720, United States. 2

Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720,

United States. 3

Department of Chemistry, University of California, Berkeley, CA 94720, United States.

4

Department of Materials Science and Engineering, University of California, Berkeley, CA

94720, United States. 5

Kavli Energy NanoScience Institute, Berkeley, California 94720, United States.

†These authors contributed equally to this work. *Correspondence to: [email protected] KEYWORDS: PbS quantum dot, nanocrystal, in-gap states, molecular oxygen, defect, scanning probe microscopy, DFT calculation

ACS Paragon Plus Environment

1

ACS Nano

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 29

ABSTRACT: Artificial solids composed of semiconductor quantum dots (QDs) are being developed for large area electronic and optoelectronic applications, but these materials often have defect-induced in-gap states (IGS) of unknown chemical origin. Here we performed scanning probe based spectroscopic analysis and density functional theory (DFT) calculations, to determine the nature of such states and their electronic structure. We found that IGS near the valence band occur frequently in the QDs except when treated with reducing agents. Calculations on various possible defects and chemical spectroscopy revealed that molecular oxygen is most likely at the origin of these IGS. We expect this impurity-induced deep IGS to be a common occurrence in ionic semiconductors, where the intrinsic vacancy defects either do not produce IGS or produce shallow states near band edges. Ionic QDs with surface passivation to block impurity adsorption are thus ideal for high efficiency optoelectronic device applications.

The evolution of semiconductor electronics and optoelectronics began with the controlled doping of impurity atoms into semiconductor crystals, which opened the way to engineer their charge transport properties.1 On the other hand, unwanted impurities or defects can hinder transport and limit device performance.2 The development of nanomaterials of controlled size made it possible to engineer the band gap and tune the electronic properties from the bottom up. As a trade-off, the impact of surface impurities on the electronic properties of nanomaterials is more prominent due to the large surface-to-volume ratio and strong quantum confinement effects.3–5 For colloidally synthesized nanocrystals or quantum dots in particular, surface-related IGS are believed to assist the recombination of photoexcited charge carriers,6 which limits the optoelectronic device efficiency.7,8 Here we focus on PbS QDs which have produced the bestperforming solar cells.9–11 Recently, Bawendi et al. reported the presence of emissive IGS that

ACS Paragon Plus Environment

2

Page 3 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

form an “effective band” inside the bandgap and ~0.2 eV away from the band edge, which were believed to be at the origin of the high open-circuit voltage deficit in the QD solar cells.12 Remarkably, this is in good agreement with our previous work where we found a band of IGS ~0.2 eV above the valence band of PbS QDs, which form percolation pathways assisting carrier transport.13,14 In addition, Sargent et al. and Klimov et al. found similar IGS using transient absorption and scanning tunneling spectroscopy methods.7,8,15 Although the existence and the energy level of IGS have been established, until now the nature of the IGS is still unclear. Theoretical predictions include: non-stoichiometry of the constituent atoms,16 incomplete ligand passivation,8 and charge-induced surface reconstruction.17 The bottleneck for the understanding of IGS is the complicated surface chemistry and ligand passivation structure of the QDs, which makes it difficult to correlate theory with experimental investigations. We overcame this obstacle with a combined effort of state-of-the-art DFT calculations and chemical spectroscopy on real PbS QD systems consisting of over a thousand atoms, and successfully identified the surface ligand passivation structure of the QDs.18 Here we extend this approach to investigate the chemical origin of IGS in PbS QDs. One experimental challenge is the accurate measurement of the energy levels of these defect states. Traditional methods include transient photovoltage measurements, thermal admittance spectroscopy, and deep level transient spectroscopy.8,19,20 These techniques provide quantitative information on the concentration of IGS, but cannot determine their energy levels. Photoluminescence, transient absorption, and photocurrent spectroscopies can probe the optical transitions in QD solids,4,7,8,12,20,21 but the correlation of the various transitions with IGS is not straightforward, considering the presence of multiple intra-band transitions and vibrational absorption of the organic ligands.22 X-ray spectroscopy techniques can probe the energy level of

ACS Paragon Plus Environment

3

ACS Nano

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 29

electronic states, but the resolution is not high enough to resolve individual IGS peaks.23–25 In fact, these spectroscopy results obtained by different groups on similar PbS QD systems vary by several 100 meV.24,25 We overcome this obstacle by using Kelvin probe force microscopy (KPFM) based surface potential spectroscopy (SPS)26 to directly resolve the energetic position of the IGS. Correlation with DFT calculations and X-ray photoelectron spectroscopy (XPS) allow us to identify the source of IGS. Previously KPFM has been used to image the local work function and built-in potential in various colloidal nanomaterials.27 However, KPFM based spectroscopic study on the density of states of QDs has not been reported before. Our methodology for studying defects in QDs, with more structural variability than the bulk single crystals explored by traditional surface science, is generally applicable to solution processed materials, such as organic thin films and perovskite materials.

RESULTS AND DISCUSSION Experimental Observation of In-Gap States. The PbS QDs studied here have an average diameter of 5.5 nm with ~0.9 eV band gap. As-synthesized colloidal QDs are typically capped with long insulating ligands.18 For lead chalcogenide (PbX, X=S, Se) QDs, various ligand exchange processes have been explored to reduce inter-particle distance and enable electron conduction in a QD solid. Here we explore treatments with 1,2-ethanedithiol (EDT), 3mercaptopropionic acid (MPA) and hydrazine (HYD) that have been reported to produce ambipolar QD field effect transistors (FETs). The resulting QDs are labeled PbS-EDT, PbSMPA and PbS-HYD, respectively.

ACS Paragon Plus Environment

4

Page 5 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

To study the IGS, in our previous work we performed scanning tunneling spectroscopy on individual PbS-EDT QDs deposited on indium tin oxide (ITO) substrate to obtain their electronic density of states (DOS).13 A heterogeneous distribution of DOS was observed on measurements over tens of QDs, with about half of them showing IGS. These states are located 0.05–0.1 eV above the Fermi level and 0.15–0.3 eV above the 1Sh states (valence band). The actual percentage of QDs with IGS can be higher since STS is only sensitive to states localized on the top surface of the QD, or with wave functions that extend all the way to the surface. The energy position of the 1Sh states was found to fluctuate from particle to particle, but compared to the 1Se states (conduction band) they are always closer to the Fermi level. This is in agreement with the previously observed p-type doping of EDT treated lead chalcogenide QDs.28,29 Here we focus on another way to measure the energy level of IGS, KPFM,26,30 which probes the change of surface potential as a function of gate bias in a QD monolayer FET. Since the occupation of states is electrostatically tuned by the gate bias in the FET, we can use KPFM to extract the density of states. In a simple electrostatic model based on the planar capacitor geometry made by the gate and the QD monolayer, the charge density in the QDs can be obtained from the measured surface potential relative to the gate voltage (Supporting Information (SI) 2.3). This spectroscopy method (SPS) has previously been used to analyze the IGS in organic thin films.26,31,32 Since the lateral resolution of our KPFM setup is 20–50 nm using a normal metal-coated tip (not the QD-coated tip previously used for ~10 nm resolution imaging13), the measured DOS is an average over tens of QDs and thus represents the ensemble properties of the QD solid, in contrast with STS that measures individual QDs. Therefore in QD systems that are typically heterogeneous in their surface chemistry and in their electronic structure, this technique provides a useful method to characterize the overall change of electronic

ACS Paragon Plus Environment

5

ACS Nano

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 29

structure with different surface chemical treatments. We can see that the DOS obtained by SPS accurately resolves the band edge of the 1Sh states and IGS of PbS-EDT (Figure 1A). The IGS is ~0.2 eV higher than the 1Sh states, in agreement with our previous STS results.13 We are unable to probe the DOS beyond the 1Sh/IGS band edge due to the limited charge injection capability of the FET setup (SI 2.3). This points to the main difference between the SPS technique and the STS technique: the former is based on electrostatic interactions and only probes the states that can be statically occupied and emptied, while the later is based on the quantum tunneling of electrons and probes a large range of states that electrons can tunnel into / out of.13,15,33,34 It has been reported that MPA passivates the QD surface more completely than EDT.7,8,21 Our SPS results on PbS-MPA (Figure 1B), however, reveals the presence of IGS at nearly the same level as those in the PbS-EDT, indicating that the IGS are not due to incomplete ligand passivation. We further performed SPS on PbS-HYD (Figure 1C), where the hydrazine acts as a reducing agent that was previously assumed to be able to repair surface defects of lead chalcogenide QDs.35,36 Note that the HYD treatment was performed directly on the original oleate coated QDs,35 instead of on the ligand exchanged QDs as reported before.37 We found that PbS-HYD has a clean band gap with no observable IGS. This result suggests that the previously observed IGS may be related with adsorbed species from the environment that are removed by the hydrazine treatment. To test this assumption, we exposed the PbS-HYD QDs to air for 30 seconds and found that IGS reappeared at ~0.1 eV above the 1Sh states (Figure 1D). This suggests that either oxygen or water is at the origin of the observed IGS. As we show next by theoretical analysis and chemical spectroscopy, O2 is the responsible species. Note that the energy level of the 1Sh states of all the samples are measured to be at -5.1 − -5.0 eV (with respect

ACS Paragon Plus Environment

6

Page 7 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

to the vacuum level), in agreement with previous results for 5.5 nm PbS QDs,38 further confirming the assignment of all the band edge states. Theoretical Calculations on In-Gap States. In our DFT calculations we used PbS QD of 4 nm diameter, for which we obtained a bandgap of 0.94 eV. This value is slightly smaller than the experimentally observed 1.1 eV bandgap for 4 nm PbS QDs,38 due to the well-known bandgap underestimation in DFT theory.39 Nevertheless, since the calculated bandgap matches that of the 5.5 nm PbS QD used in our experiments, we expect the calculated DOS to also match our experimental observations. We subsequently calculated the formation energy and electronic structure of QDs with various possible vacancies and impurities. Since the QDs are synthesized in solution, we need to consider both neutral vacancies and charged ionic vacancies (with organic counter ions in the nearby solution environment). We found that neutral atomic vacancies (Pb, S, L, where L refers to ligand), neutral divacancies (PbS, PbL2) and charged ion vacancies (Pb2+, S2-, L-) have formation energies in the range of 1.6–2.3 eV, 0.8–1.8 eV and 0.8–1 eV, respectively. Thus the formation of neutral atomic vacancies during QD growth in solution is very unlikely, while charged ion vacancies are most likely to form. Our calculations show that PbS and S2- vacancies produce IGS near the conduction band, while all other charged ion vacancies and neutral divacancies do not induce IGS (Figure S6). This leads us to conclude that external molecules adsorbed on the QD surface are at the origin of the observed IGS. These molecules need to be small and reactive in order to adsorb on the QD surface after diffusing through the ligands. Prime candidates are water and oxygen because they are present (in trace amounts) during synthesis and during glovebox storage. The calculations show that adsorbed water does not produce IGS (Figure S7), while adsorbed molecular oxygen produces IGS slightly above the valence band. Using the spin-polarized

ACS Paragon Plus Environment

7

ACS Nano

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 29

generalized gradient approximation (GGA), we found two unoccupied states at 0.07 eV and 0.18 eV above the 1Sh state (Figure 2). Higher level hybrid functional imposes a 0.13 eV correction on these states, shifting them up to 0.20 eV and 0.31 eV above 1Sh (SI 3.8). The energies of these states match those of the experimentally observed IGS in PbS-EDT, PbS-MPA, and the 30 s air exposed PbS-HYD. The two empty O2-derived IGS are formed by coupling of the two unoccupied  ∗ orbitals of molecular oxygen with the 1Sh state of the QD. The wave function density of the two IGS extend deep into the QD (Figure 2A). At the center, the wave function density still has a finite value that is 1.07% of the maximum density inside the QD. Accordingly to the Bader charge analysis of the total charge density, a partial electron transfer of 0.33 eoccurs from the QD to the oxygen molecule. The O-O distance in the adsorbed O2 is 1.278 Å, a slight increase compared to the free molecule value (1.241 Å). The calculated formation energy of the oxygen on the QD surface is -0.32 eV (Table 1), in contrast with the positive formation energies of various vacancy defects. Since the van der Waals interaction was not taken into account in the calculation, the actual formation energy can be lower (larger in absolute value) than -0.32 eV. The surface ligand species and their distribution may also affect the absolute values of the formation energy. However, we expect the relative formation energy difference between different adsorption species to be reasonable. We calculated also the DOS when oxygen molecules are dissociated on the QD surface (in minimum energy configuration), and found that no IGS are created in that case (Figure S8). Furthermore, substitutional atomic oxygen defects (where an oxygen atom replaces a sulfur atom, both at the surface and interior) were found to not produce IGS either. Therefore, among the theoretically calculated intrinsic and external defects, only molecular oxygen produces IGS with energy levels matching the experimental results.

ACS Paragon Plus Environment

8

Page 9 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

Spectroscopic Evidence and Further Theoretical Support on the Origin of IGS. To confirm the existence and analyze the concentration of molecular oxygen, we performed X-ray photoelectron spectroscopy, targeting the O 1s emission peak (Figure 3). All the samples shown in Figure 1 were measured except PbS-MPA where the ligand contains oxygen (thus complicating the spectra). XPS confirms the existence of O2 (533.9 eV) on PbS-EDT.40 Another peak at 532.1 eV corresponds to the -COO- groups, likely arising from the residual unexchanged oleate ligands.18 In PbS-HYD QDs, in contrast, O2 species are absent. When exposed to 1 atm of 25% O2 / 75% Ar gas mixture for 30 s, the 533.9 eV O2 peak appeared on the PbS-HYD. The 531.3 eV peak corresponds to -OH groups that were present on the oleate passivated QDs18 and not fully removed by the HYD treatment. Peak area analysis shows that the concentration of O2 on PbS-EDT and on the 30 s exposed PbS-HYD is roughly similar, approximately 10 molecules per QD on average. The amount of O2 on PbS-HYD remained constant after 10 min exposure (although the amount of atomic oxygen species increased) (Figure S10). The correlation of the presence or absence of O2 with the formation of IGS, together with the theoretical results on the effect of O2 in producing the IGS, lends strong support to our conclusion on the origin of the IGS. It should be noted that the XPS spectral resolution is limited by the small amount of oxygen species present on the QDs, and the resolution on the PbS-EDT sample is slightly worse than on the PbS-HYD due to the presence of ligands in the former. Also note that the existence of ~10 O2 per QD is not contradictory to the STS results where IGS was observed on only ~half of the QDs,13 since the O2 not located under the STM tip may not be sensed by STS, as explained before. Since more than one O2 molecule can be present on each QD, we further performed calculations on the effect of two O2 molecules absorbed on two opposite {100} facets. The wave functions of

ACS Paragon Plus Environment

9

ACS Nano

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 29

the IGS induced by the two molecules are symmetric and have a very small overlap at the center (Figure 4A). The IGS wave function density at the center of the QD is 1.24% of the maximum density inside the QD, slightly larger than the case of single O2. Using GGA approximation we obtained the DOS shown in Figure 4B. Four states appear inside the band gap, with two spin-up states (0.05 eV and 0.16 eV above the 1Sh state) and two spin-down states (0.06 eV and 0.18 eV above the 1Sh state). We can see that in the case of two O2, the IGS induced by each O2 resembles that of the IGS of single O2 on the QD. Similar with the single O2 case, using the higher level hybrid functional we expect the same 0.13 eV upshift in energy in the two O2 case, resulting in four states between 0.18–0.31 eV above the 1Sh state. We thus expect that the energy level of the IGS in the case of multiple O2 on one QD will not change significantly as long as these states are not strongly coupled. With two O2 on the QD, charge transfer from QD to each of the O2 molecule estimated from Bader charge analysis decreases to 0.30 e- per molecule. This is because the QD is resistant to more electron withdrawal. As a result the formation energy increases to -0.22 eV per molecule. With more O2 on the QD, we expect a further increase of formation energy. This might be the reason why the amount of O2 saturates after the PbS-HYD was exposed to oxygen for only 30 s (as shown by XPS). The formation energies of various possible impurity defects are shown in Table 1. Molecular oxygen has the largest formation energy (smallest in absolute value) among the considered impurities, and is the only one that produces IGS. We need to note that prolonged oxidation of PbS QDs results in the formation of various oxide species such as PbSO3 and PbSO4.41 In that case the simple vacancy and impurity defect picture is not sufficient to describe the electronic structure of the QD, since the composition is significantly modified.

ACS Paragon Plus Environment

10

Page 11 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

Further Discussions. We have considered various possible defects that can form during the solution phase synthesis process, which is the first step of all the device fabrication based on colloidal QDs. Among these defects, molecular oxygen, an unintentional impurity that is hard to avoid, is found to be the most likely source of IGS near the valence band. Various post-synthetic treatments, on the other hand, can modify the amount and/or type of defects, thus changing the IGS. The hydrazine treatment shown here is one example of removing O2 and eliminating IGS. Another post-synthesis approach of electronic modification is stoichiometry control via deposition of excess lead or chalcogen.42,43 Although neutral atomic vacancies are unlikely to form during solution synthesis (as discussed before), this post-synthesis control can produce QDs with excess Pb or S (i.e., QD with neutral S or Pb vacancies). According to previous calculations8,16 and our own DFT results (not shown here), neutral Pb vacancies produce IGS near the valence band, while neutral S or L (ligand) vacancies produce IGS near the conduction band. The intrinsic defect physics of lead chalcogenide QDs can be understood in the general scheme of ionic semiconductors. Similar to methylammonium lead iodide perovskite systems,44 charge balanced (surface fully passivated by ligands) PbS QDs do not show IGS. This has been demonstrated by DFT calculations on PbS QDs passivated with oleic acid and hydroxyl groups,18 with EDT,16 and with MPA + Cl.8 All the above ligands form ionic bonds with the surface Pb2+, thus maintaining the overall stoichiometry of the Pb-rich QD.45–47 In our work, to make DFT calculations possible on a QD as large as 4 nm, we used Cl as the passivation ligand and found a clean band gap after complete passivation. As long as the ligands form ionic bonds with Pb, different ligand species are expected to have similar effects on IGS properties. This is in

ACS Paragon Plus Environment

11

ACS Nano

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 29

agreement with our experimental results where similar IGS were found on QDs with three different chemical treatments (EDT, MPA, and HYD+O2). As explained before, as-synthesized QDs in solution tend to not have neutral defects due to their large formation energy. To fabricate electronically active devices, however, the QD solids typically need to go through a ligand exchange process.7,8 which can result in a loss of ligands leading to QDs with excess cation atoms (Pb in our case). These neutral vacancy defects, if present, will induce IGS near the conduction band. However, the fact that we do not observe ntype IGS indicates that either our QDs are well passivated by ligands (i.e. there is no ligand vacancy) or the IGS induced by ligand vacancies are too close (within ~50 meV) to the conduction band to be observed. In either case, O2 remains the only experimentally observable deep IGS in typical ligand exchanged QD solids that can be detrimental to optoelectronic devices. It has previously been found that oxygen exposure can induce p-type doping in lead chalcogenide QDs,48–50 supporting the O2-induced IGS mechanism. Moreover, these QD solids with various ligand treatments (e.g. EDT, n-butylamine, MPA, etc.) exhibit unintentional p-type doping even in inert atmosphere, with carrier densities of 10 − 10 for QDs of 5–6 nm diameter.42,51,52 This doping effect is also mainly due to the IGS produced by O2, whose surface concentration on the QD is limited by the energy cost of the charge transfer from the QD to the adsorbed O2 (as explained above). From our STS results on the PbS-EDT we obtain an average Fermi level ( ) – 1Sh state ( ) separation of 0.15  (averaged over the QDs with observable IGS, since those without IGS tend to have 1Sh states far from the Fermi level and are thus electronically inactive). The estimated effective density of valence band states for our 5.5 nm PbS QD solids is  = 2.1 × 10 (see Methods). Thus at room temperature the hole ACS Paragon Plus Environment

12

Page 13 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

carrier density can be calculated as (Sze 1981):  =  exp −

 ! "#

$ = 6.4 × 10 , in

good agreement with previous results. This quantitative agreement reveals that our proposed IGS mechanism is generally applicable to lead chalcogenide QDs, where unintentional doping can be induced by surface-saturated O2.

CONCLUSION We have identified molecular oxygen as the dominant source of IGS in colloidal PbS quantum dots. The in-gap electronic structure of the QDs is immune to most of the charged ion vacancies that may form during the synthesis process. To engineer IGS-free QDs for optoelectronic applications, we need to first reduce the surface oxygen during the post-synthesis treatment process, and ideally also fully passivate the surface with ligands that can block the re-adsorption of O2 on QDs.10,11

METHODS Synthesis of PbS quantum dots (QDs) was done following a previously reported procedure.53 QD size and morphology were measured by transmission electron microscopy, while their band gap is determined by UV-Vis absorption spectroscopy. QD monolayers were prepared by spincoating colloidal QDs on a substrate followed by chemical treatments and solvent washing to obtain a monolayer with different ligands or different surface chemistry. The species of the surface ligands were characterized by Fourier transform infrared spectroscopy measurements. All of the sample preparation procedures were done in Ar glovebox with